Assessing driver acceptance of technology that reduces mobile phone use while driving: The case of mobile phone applications

Accident Analysis and Prevention 135 (2020) 105348

Contents lists available at ScienceDirect

Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap

Assessing driver acceptance of technology that reduces mobile phone use while driving: The case of mobile phone applications

T

Oscar Oviedo-Trespalaciosa,b,c,*, Oliver Brianta, Sherrie-Anne Kayea, Mark Kinga a

Centre for Accident Research and Road Safety-Queensland (CARRS-Q), Institute of Health and Biomedical Innovation (IHBI), Queensland University of Technology (QUT), Brisbane, Australia b Queensland University of Technology (QUT), Institute of Health and Biomedical Innovation (IHBI), Brisbane, Australia c Department of Industrial Engineering, Universidad del Norte, Colombia

ARTICLE INFO

ABSTRACT

Keywords: Distraction Inattention Ergonomics Smartphone Driver support systems Cellphone

The nature of the road environment requires drivers to be vigilant and attentive. Distracted driving is a primary concern, as it threatens the safety of road users. However very little research has been conducted into interventions to combat such an issue. Existing interventions such as police enforcement and legislation appear to have limited effect. The use of mobile phone applications to assist in limiting driver distraction is an alternative intervention that is currently gaining traction. With a great array of potential benefits, such as reducing road toll, these applications can be readily available to all road users. Despite the positive implications, it is vital that drivers accept the use of such a technology for the intervention to be effective. Therefore, understanding driver acceptance is an important step in implanting such applications. To assess this, the present study examines the utility of two versions of the Technology Acceptance Model (TAM), the Theory of Planned Behaviour (TPB) and the Unified Theory of Acceptance and Use of Technology (UTAUT) for understanding the acceptance of technology designed to reduce distraction. Participants were presented with two different applications and responded to questions that indicated their attitudes towards the factors included in the TAM, TPB and UTAUT, alongside their intent to use the technology. A total of 731 participants responded to the survey, and their responses analysed. The results indicated that overall, Davis’ (1985) TAM was slightly better in explaining behavioural intent for both Mobile Phone Application (MPA) 1 and MPA 2, explaining 66.1% and 68.7% of the variance, respectively. Davis’ (1989) TAM and the TPB were close behind, while the UTAUT explained the least variance in behavioural intent of all the models. Overall, the findings of this study provide support for using psychological theories to assess the acceptance of mobile phone applications.

1. Introduction 1.1. Background The road environment requires drivers to be constantly vigilant to allow safe navigation around other road users, vehicles and infrastructure. Engagement in tasks that do not relate to the primary action of driving (i.e. mobile phone distraction, commercial roadside advertising, etc.) often impede this required vigilance, which increases the risk of crash and injury (Oviedo-Trespalacios et al., 2016). Specifically, mobile phone use while driving has been recognised as one of the greatest challenges for road safety today, due to the increased risk of road crashes and the high prevalence of mobile phone use. Distraction is an important issue, as numerous studies in the cognitive psychology literature have shown that dual-tasking or multitasking has important ⁎

and long-lasting perceptual and cognitive costs (Di Lollo et al., 2005; Monsell, 2003; Visser et al., 2004). The burden on attentional resources could impair driving performance for a significant period before they can be recovered. A recent naturalistic study conducted in the U.S. found that visual-manual mobile phone interactions, such as texting, increased the odds of a road traffic crash by 6.1, and dialling the phone by 12.2 (Dingus et al., 2016). In addition, the Australian Naturalistic Driving Study (ANDS) provides findings that demonstrate that Australian drivers engage in a secondary task every 1.6 min (Young et al., 2019). A systematic review of roadside distraction surveys indicates that mobile phone use while driving has been increasing around the world (Huemer et al., 2018). Despite the serious implications of distracted driving being clear, very few interventions have been found to be effective in reducing the propensity for distractions while driving. Approaches to reduce or

Corresponding author at: Queensland University of Technology (QUT), 130 Victoria Park Rd, Kelvin Grove, Queensland, 4059, Australia. E-mail addresses: [email protected], [email protected] (O. Oviedo-Trespalacios).

https://doi.org/10.1016/j.aap.2019.105348 Received 16 January 2019; Received in revised form 9 October 2019; Accepted 22 October 2019 0001-4575/ © 2019 Elsevier Ltd. All rights reserved.

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

prevent distraction during driving have included education campaigns, legislation and police enforcement. In most of the cases, these interventions have been found to have limited effectiveness. For example, Oviedo-Trespalacios (2018) found that drivers can successfully conceal their phones to avoid police enforcement. It is clear that innovative methods are required to reduce the risks relating to distraction, however there is little information to guide new strategies and interventions. The use of applications designed to reduce mobile phone distractions while driving is a potential countermeasure that has gained public attention in recent years. These applications, considered voluntary as they can be activated and deactivated at any time, limit manual and attention consuming interactions with the mobile device while the vehicle is in motion (Oviedo-Trespalacios et al., 2019a). The majority of applications currently developed focus primarily on blocking text messaging or browsing. For example, all iOS devices now come preinstalled with a ‘Do Not Disturb While Driving’ function, which disables phone features such as texting and browsing once the setting has been activated and it can detect that the vehicle is moving. Android has similarly released a function called ‘Android Auto’, which functions in much the same way. Unfortunately, applications blocking text messaging or browsing have been found to have limited popularity among younger drivers (Delgado et al., 2018). An option to overcome this limitation is to shift the mobile phone task-blocking paradigm towards the design of approaches seeking better compatibility between mobile phone tasks and driving. For example, the “Drivemode” reduces the amount of attention required to operate a mobile phone while driving through the use of large, easy to see icons, and coupled with voice commands. The application allows drivers to reply to messages through voice commands, overlay music over GPS navigation, as well as respond to or ignore calls. Authorities such as the National Highway Traffic Safety Administration (2016) have reported that the development of technologies dedicated to reducing driver workload (i.e., visual, manual, and cognitive demands) associated with performing secondary tasks is a worthwhile approach to increase safety. However, little is known about the user perspective on these mobile phone applications. Given the increasing numbers of mobile phone applications designed to prevent phone use while driving, there is ongoing need to examine public opinion and acceptance of these technologies. Acceptance of these technologies is of vital importance to the implementation of interventions for distracted driving, as an application requires usage to be effective. Despite the array of acceptance models, it is as yet unclear which most accurately describes acceptance of mobile phone applications designed for the driving context. Thus, this study will examine the utility of three psychological models in examining drivers’ acceptance of mobile phone applications.

rigorously tested in the context of acceptance of vehicle technologies and therefore, were included in the current study to examine drivers’ acceptance of mobile phone applications. Each model is briefly described below. The TAM was initially proposed by Davis (1985). In Davis’ initial model, perceived usefulness (i.e., the degree to which the system is thought to be helpful) and perceived ease of use (i.e., the degree to which the system is thought to reduce effort) were proposed to influence attitudes towards the system, which in turn influence behaviours intentions, and then actual system use (Davis, 1985). Further, Davis (1985) proposed that perceived ease of use also had a direct influence on perceived ease of use and that perceived usefulness had a direct influence on behavioural intentions. In Davis’s (1989) more recent version of the TAM, attitudes has been removed and instead, both perceived usefulness and perceived ease of use are reported to influence behavioural intentions. The TAM has been applied in the context of road safety research to assess Intelligent Transport Systems (Larue et al., 2015), vehicle navigation systems (Park et al., 2015), and more recently, to assess drivers’ acceptance of automated vehicle technology (e.g., Buckley et al., 2018; Rahman et al., 2017; Panagiotopoulos and Dimitrakopoulos, 2018). To the best of the authors’ knowledge, however, this is the first published study to apply the TAM in the context of mobile phone applications. The TPB is an additional theory which has been applied to assess acceptance in a road safety context. The TPB predicts that attitudes, subjective norms, and perceived behavioural control influence intentions, which in turn influence actual behaviour (Ajzen, 1991). In this case, attitudes refer to one’s favourable or unfavourable beliefs towards mobile phone applications, subjective norms refers to one’s perceptions that others (e.g., friends and family) would approve or disapprove of their use of mobile phone applications, and perceived behavioural control refers to the ease or difficulty of using mobile phone applications (Ajzen, 1991). While previous research has applied the TPB to assess drivers’ intentions to use hand-held mobile phones while driving (e.g., Gauld et al., 2017; Walsh et al., 2008), to date, and similar to the TAM, no research has applied this theory to assess acceptance of mobile phone applications. In addition to the TAM and the TPB, the UTAUT proposes that the factors of performance expectancy, effort expectancy, and social influence, influence behavioural intentions, which in turn, influence user behaviour (Venkatesh et al., 2003). Further, a fourth factor, facilitating condition, has been proposed to directly influence user behaviour (Venkatesh et al., 2003). Performance expectancy was defined by Venkatesh et al. (2003) as the degree to which a system will assist an individual. Effort expectancy refers to the ease of use of the system, and similar to subjective norms in the TPB, social influence refers to the extent to which important other would approve or disapprove of use (Venkatesh et al., 2003). Finally, facilitating condition refers to the degree to which an individual believes that there is support available to use the system (Venkatesh et al., 2003). In addition, Venkatesh et al. (2003) proposed that age, gender, experience, and voluntariness of use were moderating factors. Previous research has applied the UTAUT to assess drivers’ acceptance of new technologies, such as automated road transport systems (Madigan et al., 2017) and advanced driver assistance systems (Rahman et al., 2017). This research will extend upon these previous studies by applying the UTAUT to assess the new technology of mobile phone applications.

1.2. Theoretical models of technology acceptance Psychological models of acceptance can be applied to assess the extent to which mobile phone applications are accepted by drivers. This study applied three psychological models of technology acceptance, including; the Technology Acceptance Model (TAM; Davis, 1985, 1989), the Theory of Planned Behaviour (TPB; Ajzen, 1991), and the Unified Theory of Acceptance and Use of Technology (UTAUT; Venkatesh, Morris, Davis, & Davis, 2013). The TAM and UTAUT are two of the most widely applied models used to explain technology acceptance and are supported by previous research for their predictive power of behavioural intentions (e.g., De Angelis et al., 2017; Madigan et al., 2017; Rahman et al., 2017). More recently, the TPB has been applied to assess acceptance of technologies, such as advanced driver assistance systems (e.g., Rahman et al., 2017), automated vehicle technologies (e.g., Buckley et al., 2018), in-vehicle information systems (IVIS) (e.g., Oviedo-Trespalacios et al., 2019e) and intelligent transport systems (e.g., Larue et al., 2015). While it is acknowledged that other psychological models do exist, the TPB, TAM, and UTAUT have been

1.3. The current study Driver acceptance is a necessary first step for use of voluntary applications to reduce mobile phone distracted driving, however this in turn requires an understanding of the factors that contribute to driver acceptance, and how they interact. There were two overarching research aims of this study. First, the research aimed to explore and compare the explanatory value of the TAM, the TPB and the UTAUT in a 2

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

Table 1 Descriptions of Mobile Phone Application 1 and Mobile Phone Application 2. MPA 1: Do Not Disturb While Driving

MPA 2: Drivemode

Source: phonedog.com and 9to5mac.com

Source: drivemode.com Drive Mode Simplifies phone usage while driving. Utilises a streamlined interface to safely access certain applications, with voice-enabled commands and large icons to minimise distraction Reply to a message or text using voice commands Overlay music on top of a navigation app Ignore calls or messages in ‘Do Not Disturb’ mode Set up text message auto-replies Automatically launch and close with Bluetooth pairing Configure to automatically launch when driving commences View driving history for personalised recommendations over time based on activities, calendar and favourite places Integrates with many existing apps, including: ○ Google, Waze, HERE Maps ○ Pandora, Spotify, Google Play Music, Player.fm, Poweramp ○ SMS, Facebook Messenger, Slack, Whatsapp ○ Google Now and Google Assistant

be initiated voluntarily once and then will detect driving automatically. • Must be deactivated within first 5 minutes, if phone user is a passenger. • Can testing and all other mobile phone tasks that require prolonged screen • Blocks observation such as social media, texts, etc. while the vehicle is moving. not allow notifications (audio or vibration) whilst the vehicle is moving. • Does Incoming calls are allowed if the phone is not connected to Bluetooth and hands-free

• • • • • •

• • • • • • • • •

device calls, in that case the app will read who is calling so you ca decide to answer or not. Bluetooth or hands-free device calls are permitted including emergency services. Automatically responds to messages to notify sender that recipient is driving, if preferred. Records driving performance for monitoring and self-check purposes If desired, it shares performance data with family and friends when desired, in cases of collaboration to reduce distraction Allows usage of GPS and music apps Otherwise, the phone cannot be used.

sample of licensed drivers. While other studies have found that these theories can effectively explain acceptance behaviours in other contexts, very little study has been undertaken in exploring and comparing the efficacy of each model in the distracted driving context. Second, this study also aimed to determine the intent of participants to use each of the proposed applications to prevent distracted driving. These technologies are in the early stages of design and development and it is therefore necessary to establish tools to evaluate acceptance of this promising technology (Oviedo-Trespalacios et al., 2019a).

2.2. Measures In order to assess the acceptance and behavioural intent of mobile phone applications, a number of different measures were used, based on the different components of the TAM, TPB and UTAUT. The measures used in the present study mirror those utilised by Rahman et al. (2017), which are adapted from those used by Van Der Laan et al. (1997); Venkatesh and Davis (2000); Venkatesh et al. (2003) and Adell (2010). For all measures, overall scores of each construct were determined through calculating the mean score across each question. Assessment of the TAM comprised three constructs: attitude, perceived usefulness and perceived ease of use. Attitude was measured through 10 Likert-type scales. Each question asked the participant “The use of the system when I am driving would be:” with options differing, such as; ‘Bad: 1’ to ‘Good: 7’, and ‘Sleep inducing: 1’ to ‘Alerting: 7’. Assessment of perceived usefulness and perceived ease of use in the TAM, and all measures of the TPB and UTAUT, as well as the dependent variable of behavioural intention, utilised 7-point Likert-type scales, with scores ranging from 1 “Strongly Disagree” to 7 “Strongly Agree”. Perceived usefulness utilised four items (i.e. ‘Using the system would improve my driving performance’) and perceived ease of use comprised four items (i.e. I would find it easy to get the system to do what I want it to do’), the second item of which was reverse scored. Assessment of the TPB comprised two constructs: subjective norms and perceived behavioural control. Two items were used to assess subjective norms (i.e. ‘People who influence my behaviour would think that I should use the system’) and four items assessed perceived behavioural control (i.e. ‘I have the resources necessary to use the system’). The UTAUT assessment comprised three constructs: performance expectancy, effort expectancy and social influence. Facilitating condition was not measured for this study, as it does not predict behavioural intent within the UTAUT model. Four items assessed performance expectancy (i.e. ‘I would find the system useful in my driving’), four items assessed effort

2. Method 2.1. Participants Participants were recruited through advertising the survey through a media release from the Queensland University of Technology, through social media methods, and through an email sent through insurance clubs. A total of 731 participants completed the survey, of which 52.1% (n = 381) were male, 47.7% (n = 349) were female, and 0.1% (n = 1) unspecified. Ages ranged from 17 to 90 years, with a mean of 47.13 years (SD = 17.89). The majority of participants held an Australian open drivers licence (88.9%), while 7.3% held an Australian provisional drivers licence, and 3.8% held an international drivers licence. It was found that only 23.4% of participants drove a car with a manual transmission, while the remaining 76.6% reported driving an automatic car. Further, 43.4% of participants reported that their primary purpose of driving was a mixture of work/commute to work and personal, while participants that used their vehicles for mostly work/commuting to work or for mostly personal reasons comprised 21.3% and 35.3% respectively. On average, participants drove 9.38 h a week (SD = 7.92). In terms of mobile phone devices, 47.7% of participants reported using iOS devices, while 46.2% of participants reported using Android devices. Only 2.5% reported using a Windows device. 3

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

expectancy (i.e. ‘I would find the system easy to use’) and the measure of social influence was the same as that of subjective norms outlined earlier. The dependent measure of behavioural intention consisted of seven items (i.e. ‘If my phone is equipped with a similar system, I expect that I would use the system when driving”). Further, familiarity with the application was measured through a single item 7-point Likert-type scale, with scores ranging from 1 “I have never heard of a similar driving app” to 7 “I regularly use a similar system when driving”.

Table 2 Internal consistencies and descriptive statistics of the measures for MPA 1 and MPA 2. Factors

BI A PU PEoU SN/SI PBC PE EE FAM Age

2.3. Procedure and data analysis Participants completed the questionnaire online regarding their demographics (i.e. gender, age, hours driving, purpose of driving, education, etc.) and opinions about the use of technology to reduce driver distraction. Two main types of applications were tested in this study: MPA1 oriented to block mobile phone interactions while driving, based on the iOS “Do Not Disturb While Driving” feature, and MPA2 oriented to reduce mobile phone use workload through the provision of a simplified interface such as “Drivemode”. Table 1 presents a summary of the applications as presented to the participants during the study. More information on the application can be found in the comprehensive review conducted by Oviedo-Trespalacios et al. (2019a). Participants were exposed to one application at a time, presented in sequential but random order. They responded to questions regarding previous experience with similar mobile phone technology (such as familiarity with the applications), and the measures outlined in section 2.2 that indicated various beliefs about the application systems and the participants’ intent to use the system. Data analysis was conducted using descriptive statistics (i.e., means, reliability and correlations) and hierarchical multiple regressions. Each of the proposed models of the TAM, TPB and UTAUT was examined using individual hierarchical multiple regression analyses. The analysis was conducted in IBM SPSS statistics (version 23), and was guided by Rahman et al. (2017).

MPA 1

MPA 2

Mean

S.D.

α

Mean

S.D.

α

4.22 5.00 4.43 4.78 3.81 5.01 3.82 5.00 1.55 47.13

1.91 1.63 1.85 1.25 1.95 1.53 1.95 1.31 1.13 17.89

.95 .96 .93 .69 .90 .88 .94 .78 – –

3.67 4.43 3.87 4.54 3.59 4.76 4.35 4.77 1.48 47.13

1.94 1.77 1.97 1.31 1.96 1.64 1.92 1.37 1.11 17.89

.96 .98 .96 .70 .94 .94 .96 .77 – –

Note: Gender was not included in descriptive table as there were no means or standard deviations to report. Note: Cronbach’s alpha not possible to calculate for Familiarity (Single item measure) or for age. Abbreviated: BI – Behavioural Intention, A – Attitude, PU – Perceived Usefulness, PEoU – Perceived Ease of Use, SN – Subjective Norms, PBC – Perceived Behavioural Control, PE – Performance Expectancy, EE – Effort Expectancy, SI – Social Influence, FAM – Familiarity.

mobile phone distraction while driving and 89.9% of participants have never experienced driving with an MPA designed to reduce driver distraction. Further still, 71.0% report never hearing of an application similar to MPA 1, and 75.6% report never hearing of an application similar to MPA 2. 3.2. Paired samples t-test A paired samples t-test was conducted to examine the overall difference in behavioural intention between MPA 1 and MPA 2. It was found that there was a significant difference in behavioural intention between MPA 1 (M = 4.22, SD = 1.91) and MPA 2 (M = 3.67, SD = 1.94), t(730) = 8.02, p < .01, d = 0.29. These results indicate that participants reported a significantly greater intent to use MPA 1 than MPA 2.

3. Results 3.1. Descriptive statistics and reliability scales of measures for MPA 1 and MPA 2

3.3. Hierarchical multiple regressions

The reliability statistics were calculated independently for MPA 1 and MPA 2. The internal consistencies for the measures used to assess MPA 1 were found to have good reliability, with most Cronbach’s alpha scores of above .70 (Table 2). The only measure with a Cronbach’s alpha of less than .70 was the measure of perceived ease of use, which had a slightly lower score of .68. As the second item was reverse scored, it is possible that some participants were not paying attention to this scoring and answered the question as they would have the others, thus accounting for the discrepancy within scores. Despite this, as the internal consistency is very close to .70, the measure was retained as a whole. The internal reliabilities for familiarity, age or gender were unable to be calculated, due to the single-item nature of familiarity and age, and the dichotomous nature of gender. Bivariate correlations between measures of the MPA 1 are outlined in Table 3. Most variables were found to be significantly positively correlated at the 0.05 level (two tailed), with the exception of familiarity, gender and some points of age. The internal consistencies of the measures used to assess MPA 2 were all found to have good reliabilities, with all Cronbach’s alpha scores being higher than .70, aside from the single item measures of age and familiarity, and the dichotomous variable of gender once again. Bivariate correlations between MPA 2 variables are outlined in Table 4. Contrary to MPA 1, familiarity with MPA 2 and age was found to be significantly positively correlated to most measures, with the exception on performance expectancy. The means and standard deviations of each of the measure are outlined in Table 1. Prior to the completion of this study, 60.1% of participants had never heard of MPAs that reduce

3.3.1. Controlling for familiarity, age and gender In each of the models listed below, familiarity with each of the MPAs is added in step one of the hierarchical regression to control for participants’ prior experience with the application. Age and gender are similarly entered in the first step so as to control for individual variability. As the results remain consistent across all four models, only the results of the second step will be reported. The analyses of MPA 1 revealed that familiarity and age were significant contributors to the overall model of behavioural intention in MPA 1, whereas gender did not significantly predict behavioural intention (F(3, 727) = 3.30, p = .020), accounting for only 0.2% of the variance in behavioural intention. In contrast to MPA 1, the model for MPA 2 found that familiarity significantly accounted for 2.0% of the variance in behavioural intention when entered in step one, whereas gender and age were not significant predictors (F(3, 727) = 4.83, p = .002). 3.3.2. TAM (Davis, 1985) A two-step hierarchical multiple regression analysis was conducted with behavioural intention as the dependent variable. Familiarity, age and gender were entered in step one, attitude and perceived usefulness were entered in step two (Table 5). The addition of attitude and perceived usefulness in step two significantly accounted for an additional 66.3% of the variance in behavioural intention of MPA 1, above and beyond that explained by familiarity (F(2, 725) = 706.91, p < .01). 4

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

Table 3 Bi-variate correlations for MPA 1. Factors

BI

A

PU

PEoU

SN/SI

PBC

PE

EE

FAM

Age

Gender

BI A PU PEoU SN/SI PBC PE EE FAM Age Gender

– .69** .77** .54** .61** .51** .47** .48** .05 .08* .02

– .64** .52** .44** .40** .32** .40** .01 .12** .05

– .55** .65** .50** .56** .48** .05 −.04 .08*

– .36** .69** .24** .75** .07 −.04 .05

– .39** .51** .33** .02 −.05 −.01

– .30** .78** .13** −.08* .03

– .26** .04 −.13** .09*

– .12** −.10** .05

– −.18** −.04

– −.25**

–

Note: Internal consistency (Cronbach’s alpha) statistics are on the diagonal. Note: SN and SI utilised the same two items, also referred to as Attitude in the TPB. Abbreviated: BI – Behavioural Intention, A – Attitude, PU – Perceived Usefulness, PEoU – Perceived Ease of Use, SN – Subjective Norms, PBC – Perceived Behavioural Control, PE – Performance Expectancy, EE – Effort Expectancy, SI – Social Influence, FAM – Familiarity. * Correlation is significant at 0.05 level (2-tailed). ** Correlation is significant at 0.01 level (2-tailed).

Perceived usefulness and attitude were significant positive predictors, age remained significant, gender remained non-significant and familiarity became non-significant. It was found that perceived usefulness was a stronger predictor of behavioural intention than attitude or familiarity. For MPA 2, attitude and perceived usefulness were added to the model in step two and accounted for 69.2% of the variance in behavioural intention, with perceived usefulness and attitude being significant positive predictors and familiarity becoming non-significant (F (2, 725) = 799.86, p < .01). Age remained non-significant, however gender became significant. As with MPA 1, perceived usefulness was found to be a greater predictor of behavioural intention than attitude or familiarity.

or familiarity. For MPA 2, the addition of perceived usefulness and perceived ease of use in step 2 found that the model significantly accounted for 67.1% of the variance in behavioural intention, with perceived usefulness and perceived ease of use once again being significant positive predictors whereas familiarity was found to be non-significant (F(2, 725) = 725.69, p < .01). Age and gender were found to be significant. As with MPA 1, perceived usefulness was a stronger predictor of behavioural intention than perceived ease of use or familiarity. 3.3.4. TPB (Ajzen, 1991) A two-step hierarchical multiple regression was once again carried out to analyse the effects of the predictors on behavioural intention in both MPA 1 and MPA 2 (Table 7). Familiarity, age and gender were once again controlled for by being added in the first step, and attitude, subjective norm and perceived behavioural control were added in the second step. For MPA 1, the addition of attitude, subjective norm and perceived behavioural control significantly accounted for 63.3% of the variance in behavioural intention (F(3, 724) = 413.15, p < .01), where attitude, perceived usefulness and perceived ease of use were significant positive predictors. Age remained significant, gender remained non-significant, while familiarity became non-significant. It was found that attitude was a stronger predictor of behavioural intention than subjective norm or perceived behavioural control. When examining MPA 2, the results found that the addition of attitude, subjective norm and perceived behavioural control in the second step significantly accounted for 67.0% of the variance in behavioural

3.3.3. TAM (Davis, 1985, 1989) A two-step hierarchical multiple regression was once again carried out to analyse the effects of the predictors on behavioural intention in both MPA 1 and MPA 2 (Table 6). Familiarity, gender and age were once again controlled for by being added in the first step, and perceived usefulness and perceived ease of use were added in the second step. For MPA 1, the addition of perceived usefulness and perceived ease of use in the second step significantly accounted for 62.3% of the variance in behavioural intention (F(2, 725) = 593.50, p < .01), with attitude and perceived usefulness being significant positive predictors, gender remained non-significant, age remained significant and familiarity became non-significant. It was found that perceived usefulness was a stronger predictor of behavioural intention than perceived ease of use Table 4 Bi-variate correlations for MPA 2. Factors

BI

A

PU

PEoU

SN/SI

PBC

PE

EE

FAM

Age

Gender

BI A PU PEoU SN/SI PBC PE EE FAM Age Gender

– .72** .81** .49** .71** .49** .50** .41** .13** −.08* .02

– .75** .47** .57** .45** .38** .41** .15** −.15** .12**

– .50** .76** .50** .62** .43** .14** −.15** .10**

– .36** .79** .27** .85** .19** −.23** .09*

– .40** .55** .32** .09* −.12** .04

– .31** .36** .17** −.15** .05

– .25** −.03 −.03 .05

– .16** −.19** .06

– −.19** −.02

– −.25**

–

Note: SN and SI utilised the same two items, also referred to as Attitude in the TPB. Abbreviated: BI – Behavioural Intention, A – Attitude, PU – Perceived Usefulness, PEoU – Perceived Ease of Use, SN – Subjective Norms, PBC – Perceived Beahvioural Control, PE – Performance Expectancy, EE – Effort Expectancy, SI – Social Influence, FAM – Familiarity. * Correlation is significant at 0.05 level (2-tailed). ** Correlation is significant at 0.01 level (2-tailed). 5

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

Table 5 Assessment of the Technology Acceptance Model (Davis, 1985). Steps

MPA 1

MPA 2

2

Adj. R 1. Control Behavioural Intention Familiarity Gender (Female) Age 2. BI = A + PU + FAM Behavioural Intention Attitude Perceived Usefulness Familiarity Gender (Female) Age

0.01

0.66

B

SE B

95% CI

ß

0.13 0.20 0.01

0.06 0.15 0.01

0.00, 0.25 −0.09, 0.48 0.00, 0.02

0.08* 0.05 0.11**

0.40 0.58 0.06 −0.07 0.01

0.03 0.03 0.04 0.09 0.01

0.31, 0.44 0.53, 0.64 −0.01, 0.13 −0.24, 0.10 0.00, 0.01

0.32** 0.56** 0.04 −0.02 0.06**

Adj. R2 0.02

0.69

B

SE B

95% CI

ß

0.21 0.02 −0.01

0.07 0.15 0.01

0.08, 0.34 −0.27, 0.31 −0.01, 0.00

0.12** 0.01 −0.05

0.30 0.61 0.02 −0.26 0.01

0.03 0.03 0.04 0.08 0.01

0.23, 0.37 0.55, 0.67 −0.05, 0.09 −0.43, -0.10 0.00, -0.08

0.27** 0.62** 0.01 −0.07** 0.04

* p < 0.05. ** p < 0.01.

intention (F(3, 724) = 481.80, p < .01) where perceived usefulness and perceived ease of use were significant positive predictors, while familiarity was non-significant. Both gender and age became significant in the second step. As with MPA 1, attitude was a stronger predictor of behavioural intention than subjective norm, perceived behavioural control or familiarity.

3.4. Comparison between TAM, TPB and UTAUT

3.3.5. The UTAUT (Venkatesh et al., 2003) A two-step hierarchical multiple regression was once again carried out to analyse the effects of the predictors on behavioural intention in both MPA 1 and MPA 2 (Table 8). Familiarity, age and gender were once again controlled for by being added in the first step, and performance expectancy, effort expectancy and social influence were added in the second step. For MPA 1, the addition of the predictors in the second step significantly accounted for 50.2%% of the variance in behavioural intention (F(3, 724) = 240.88, p < .01). Within this model, it was found that social influence was the highest predictor of behavioural intention, followed very closely by effort expectancy, both of which were greater than the effect of performance expectancy and familiarity. In relation to MPA 2, the results demonstrated that the addition of the predictors at step 2 accounted for 56.1% of the variance in behavioural intention (F(3, 724) = 302.38, p < .01) with performance expectancy, effort expectancy and social influence being significant predictors. Familiarity was similarly significant, however age and gender were found to be non-significant. Within this model, it was found that social influence was a greater predictor of behavioural intention than performance expectancy, effort expectancy or familiarity.

4. Discussion

Through observing the Beta scores listed above for each model, comparisons can be drawn between each model (Fig. 1). It can be seen that for the TAM model (Davis, 1985) accounts for the greatest variance in behavioural intention across both MPA 1 and MPA 2.

This study utilised a cross-sectional design to investigate driver acceptance of mobile phone applications to prevent distracted driving. Mobile phone applications to prevent distracted driving are considered an innovative and promising strategy to reduce distraction (OviedoTrespalacios et al., 2019a; Parnell et al., 2018). Two main types of applications were tested in this study: an application oriented to block mobile phone interactions while driving, based on the iOS “Do Not Disturb While Driving” feature, and an application oriented to reduce mobile phone use workload through the provision of a simplified interface such as “Drivemode”. As explained by Oviedo-Trespalacios et al. (2019a), these are the two fundamental approaches for the design of mobile phone applications to prevent distracted driving available today. In practice, these applications have the potential to reduce exposure to risky mobile phone such as texting or browsing, which is extremely positive in terms of safety. However, there is little evidence regarding the acceptability of mobile phone applications by drivers who use their phones while driving (Delgado et al., 2018; McGinn, 2014; Oviedo-Trespalacios et al., 2019b). Consequently, to be considered an

Table 6 Assessment of the Technology Acceptance Model (Davis, 1985, 1989). Steps

MPA 1 Adj. R2

1. Control Behavioural Intention Familiarity Gender (Female) Age 2. BI = PU + PEoU Behavioural Intention Perceived Usefulness Perceived Ease of Use Familiarity Gender (Female) Age

0.01

0.62

MPA 2

B

SE B

95% CI

ß

0.13 0.20 0.01

0.06 0.15 0.01

0.00, 0.25 −0.09, 0.48 0.00, 0.02

0.08* 0.05 0.11**

0.71 0.25 0.05 −0.03 0.01

0.03 0.04 0.04 0.09 0.01

0.65, 0.76 0.17, 0.34 −0.03, 0.13 −0.21, 0.15 0.01, 0.02

0.70** 0.17** 0.03 −0.01 0.12**

* p < 0.05. ** p < 0.01. 6

Adj. R2 0.02

0.67

B

SE B

95% CI

ß

0.21 0.02 −0.01

0.07 0.15 0.01

0.08, 0.34 −0.27, 0.31 −0.01, 0.00

0.12** 0.01 −0.05

0.75 0.19 0.02 −0.22 0.01

0.02 0.04 0.04 0.09 0.01

0.70, 0.79 0.11, 0.26 −0.05, 0.10 −0.39, -0.06 0.00, 0.01

0.76** 0.13** 0.01 −0.06** 0.05*

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

Table 7 Assessment of the TPB (Ajzen, 1991). Steps

MPA 1 Adj. R

1. Control Behavioural Intention Familiarity Gender (Female) Age

2

0.01

2. BI = A + SN + PBC + FAM Behavioural Intention 0.63 Attitude Subjective Norm Perceived Behavioural Control Familiarity Gender (Female) Age

MPA 2

B

SE B

95% CI

ß

Adj. R

0.13 0.20 0.01

0.06 0.15 0.01

0.00, 0.25 −0.09, 0.48 0.00, 0.02

0.08* 0.05 0.11**

0.54 0.33 0.25

0.03 0.03 0.03

0.47, 0.60 0.28, 0.38 0.19, 0.32

0.46** 0.34** 0.20**

0.05 0.06 0.01

0.04 0.09 0.01

−0.03, 0.13 −0.11, 0.24 0.00, 001

0.03 0.02 0.06**

2

0.02

0.67

B

SE B

95% CI

ß

0.21 0.02 −0.01

0.07 0.15 0.01

0.08, 0.34 −0.27, 0.31 −0.01, 0.00

0.12** 0.01 −0.05

0.48 0.41 0.15

0.03 0.03 0.03

0.42, 0.54 0.36, 0.47 0.09, 0.21

0.44** 0.42** 0.13**

0.03 −0.19 0.01

0.04 0.09 0.01

−0.05, 0.10 −0.36, -0.02 0.00, 0.01

0.02 −0.05* 0.05*

* p < 0.05. ** p < 0.01.

effective countermeasure, it is necessary to gain an understanding of the potential factors influencing acceptability of mobile phone applications to prevent distracted driving. This study found significant differences in the behavioural intentions to use the applications between MPA 1 and MPA 2. Participants reported higher behavioural intentions to use MPA 1, which seeks to block mobile phone interactions compared to MPA 2 which seeks to reduce workload of using a mobile phone. This finding suggests that drivers consider the differences in the design approach as a critical factor to define their intention to use the mobile phone applications and currently prefer the blocking approach. This can be result of the extensive efforts occurring in Australia to prevent phone use while driving by increasing risk perception and increasing negative attitudes in the community. Arguably, MPA 2: drivemode represents a departure from traditional cessation approaches. The implications for crash prevention are enormous given that it has been demonstrated that blocking apps (i.e., MPA 1: “Do Not disturb While Driving”) could reduce exposure to risky mobile phone interactions (Albert and Lotan, 2019; Ponte et al., 2016; Oviedo-Trespalacios et al., 2019c). Unfortunately, the adoption of these applications has been low among drivers in countries such as Australia (Oviedo-Trespalacios et al., 2019b) and the U.S (Reagan and Cicchino, 2018). More efforts are needed to increase acceptance and knowledge of this technology among drivers.

Another important implication of the relatively lower intention to use MPA 2: Drivemode is that user acceptance might be a challenge in the development of technologies seeking the safe integration between mobile phones. The development of a simplified interface to reduce workload and support safe co-occurrence of distraction and driving, as depicted by the MPA2, has been a recommend approach to deal with distraction by authorities in the U.S. (National Highway Traffic Safety Administration [NHTSA], 2016). Specifically, the NHTSA’s Driver Distraction Guidelines suggest limiting the visual, manual, and cognitive demand associated with secondary tasks performed using these devices to minimise the time and effort involved in a driver performing a task using a particular device. In the case of mobile phone use while driving, reducing workload of mobile phone tasks could be a more sensible approach, given that mobile technologies have an increasing role in our lifestyle, including professional, social, and academic contexts to the point that experts have agreed that stopping using them is nearly impossible (Panova and Carbonell, 2018). As such, more specialized designs capable of increasing end-user acceptance are required to support the development of mobile phone applications to prevent distracted driving. The findings show support for applying theoretical models of technology acceptance in the context of mobile phone applications. While each of the three models was able to predict intentions to use the two

Table 8 Assessment of the UTAUT (Venkatesh et al., 2003). Steps

MPA 1 Adj. R

1. Control Behavioural Intention Familiarity Gender (Female) Age 2. BI = PE + EE + SI Behavioural Intention Performance Expectancy Effort Expectancy Social Influence Familiarity Gender (Female) Age

0.01

0.50

2

MPA 2

B

SE B

95% CI

ß

0.13 0.20 0.01

0.06 0.15 0.01

0.00, 0.25 −0.09, 0.48 0.00, 0.02

0.08* 0.05 0.11**

0.19 0.43 0.42 0.06 0.15 0.02

0.03 0.04 0.03 0.05 0.10 0.01

0.13, 0.25 0.35, 0.51 0.36, 0.48 −0.03, 0.15 −0.05, 0.35 0.01, 0.02

0.19** 0.30** 0.43** 0.04 0.04 0.17**

* p < 0.05. ** p < 0.001. 7

Adj. R 0.02

0.56

2

B

SE B

95% CI

ß

0.21 0.02 −0.01

0.07 0.15 0.01

0.08, 0.34 −0.27, 0.31 −0.01, 0.00

0.12** 0.01 −0.05

0.14 0.28 0.57 0.11 −0.06 0.01

0.03 0.04 0.03 0.04 0.10 0.01

0.08, 0.20 0.20, 0.35 0.51, 0.63 0.02, 0.20 −0.25, 0.13 −0.00, 0.01

0.14** 0.20** 0.58** 0.06* −0.02 0.05

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

Fig. 1. Comparison of Models for MPA 1 and MPA 2 on Behavioural Intention.

mobile phone applications, after controlling for familiarity, age, and gender, the TAM factors were shown to explain more variance in behavioural intent for MPA 1 and MPA 2 when compared to the TPB and UTAUT. Both the TAM and TPB factors explained above 60% of variance in behaviour intentions to use the two mobile phone applications, the UTAUT factors explained 50% and 56% in MPA 1 and MPA 2, respectively. Overall, the findings show that the TAM and the TPB may be more suited to explain acceptance of, and intentions to use, mobile phone applications when compared to the UTAUT. However, given the limited research on applying theoretical models of technology acceptance in the context of mobile application use designed to prevent distracted driving, further research is required to assess the utility of the TAM, TPB, and UTAUT. Nonetheless, the levels of explained variance reached in this study are relatively superior to previous studies predicting intentions of using in-vehicle systems such as GPS navigation (41.5%) (Yao et al., 2019). Given the suitability of the TPB and TAM to predict intentions, these theories could inform interventions to reduce mobile phone use while driving. Attitudes and perceived usefulness of the application, as defined in the TPB and TAM respectively, were the most important predictors of intention to use an application to prevent mobile phone use while driving. Therefore, both constructs could be targeted and improved by appropriate educational methods among motorists to increase the adoption of this technology. Previous research has shown that intervention based on psycho-social frameworks such as the TPB can be effective in behaviour change interventions (Steinmetz et al., 2016). Considering the low uptake of these mobile phone applications (Reagan et al., 2018; Oviedo-Trespalacios et al., 2019b), an important lesson of this study is that the successful adoption of driver safety technology must go beyond than simply installing a solution or product. Future research is needed to develop appropriate countermeasures, but the applicability of psychosocial frameworks such as TPB and TAM is promising. Regarding the demographic variables, women seem to have a lower intention to use the applications compared to males. This is very concerning given that recent research has consistently suggested that women are more likely to engage in mobile phone use while driving at any time (Oviedo-Trespalacios et al., 2018). This finding further supports the assumption that gender-related factors play an important role in the success of road safety interventions. Future research needs to be directed to understand how these applications can overcome potential gender-specific barriers for the intake of new technology. Additionally, the results show that age has an important impact on the intention of using a blocking app (i.e., MPA 1: “Do Not disturb While Driving”) compared to a mobile phone application oriented to

reduce workload. This result can be explained by the fact that young people today might experience more difficulty self-regulating their mobile phone use compared to older generations. Research in Australia has shown that young people report higher levels of problematic mobile phone use, which refers to an individual's inability to control their usage of their mobile phone resulting in adverse consequences (OviedoTrespalacios et al., 2019d). This means that an application that still enables some degree of interaction with a mobile phone while driving would be accepted by a more age-diverse group of drivers. An important consideration here is that mobile phone use while driving is generally widespread among drivers of all ages in alarming rates, with young drivers being at a higher risk (Oviedo-Trespalacios et al., 2017). Therefore, approaches for application design that can appeal to a larger group of the population are desirable. A number of methodological limitations need to be emphasized. Firstly, the applications utilised in this research might not be the most adequate to reduce crash risk due to mobile phone use. As explained by Oviedo-Trespalacios et al. (2019a), mobile phone applications that facilitate safer mobile phone use while driving are in the early stages of design and development and very little is known about their benefits in terms of crash risk. Although the mobile phone applications tested in this study are commercially available, we cannot endorse the use of any one application over another without an assessment of crash risk. Future research is needed to study the acceptance of more advanced technologies. Secondly, despite the survey being completely anonymous and voluntary in an attempt to minimise the issues associated with self-report data, it is possible that participants may have been influenced by social desirability biases and recorded inaccurate responses. Thirdly, sample recruitment was not random and could also have impacted the results. Therefore, the sample might not be representative. Future research should consider the use of observational methods to study acceptance. Overall, the findings have provided support for applying theoretical models of technology acceptance in the context of mobile phone applications. Given that distracted driving has been shown to lead to death and serious injuries on our roads, new technologies such as mobile phone applications, are vital in order to prevent distracted driving. Applying well validated models, such as the TAM, TPB, and UTAUT, may assist in increasing our understanding of the extent to which these new technologies are accepted by the community, and in turn, may assist in preventing crashes associated with mobile phone use. Declaration of Competing Interest The authors declare that they have no known competing financial 8

Accident Analysis and Prevention 135 (2020) 105348

O. Oviedo-Trespalacios, et al.

interests or personal relationships that could have appeared to influence the work reported in this paper.

mobile phone use while driving in Queensland: prevalence, attitudes, crash risk perception, and task-management strategies. PLoS One 12 (9), e0183361. Oviedo-Trespalacios, O., Haque, M.M., King, M., Washington, S., 2018. Should I text or call Here? A situation-based analysis of drivers’ perceived likelihood of engaging in mobile phone multitasking. Risk Anal. 38 (10), 2144–2160. https://doi.org/10.1111/ risa.13119. Oviedo-Trespalacios, O., King, M., Vaezipur, A., Truelove, V., 2019a. Can our phones keep us safe? A content analysis of smartphone applications to prevent mobile phone distracted driving. Transp. Res. Part F Traffic Psychol. Behav. 657–668. https://doi. org/10.1016/j.trf.2018.11.017. Oviedo-Trespalacios, O., Williamson, A., King, M., 2019b. User preferences and design recommendations for voluntary smartphone applications to prevent distracted driving. Transp. Res. Part F Traffic Psychol. Behav. 64, 47–57. https://doi.org/10. 1016/j.trf.2019.04.018. Oviedo-Trespalacios, O., King, M., Truelove, V., Kelly, R., 2019c. Can Voluntary Apps Reduce Mobile Phone Use While Driving? RACV. Retrieved from https://www.racv. com.au/content/dam/racv/images/public-policy/reports/RACV_Voluntary%20Apps %20Report_WEB.pdf. Oviedo-Trespalacios, O., Nandavar, S., Newton, J.D.A., Demant, D., Phillips, J.G., 2019d. Problematic use of mobile phones in Australia… is it getting worse? Front. Psychiatry 10 (105), 1–15. https://doi.org/10.3389/fpsyt.2019.00105. Oviedo-Trespalacios, O., Nandavar, S., Haworth, N., 2019e. How do perceptions of risk and other psychological factors influence the use of in-vehicle information systems (IVIS)? Transp. Res. Part F Traffic Psychol. Behav. 67, 113–122. https://doi.org/10. 1016/j.trf.2019.10.011. Panagiotopoulos, I., Dimitrakopoulos, G., 2018. An empirical investigation on consumers’ intentions towards autonomous driving. Transp. Res. Part C Emerg. Technol. 95, 773–784. https://doi.org/10.1016/j.trc.2018.08.013. Panova, T., Carbonell, X., 2018. Is smartphone addiction really an addiction? J. Behav. Addict. 7 (2), 252–259. Park, E., Kim, H., Ohm, J.Y., 2015. Understanding driver adoption of car navigation systems using the extended technology acceptance model. Behav. Inf. Technol. 34 (7), 741–751. Parnell, K.J., Stanton, N.A., Plant, K.L., 2018. Creating the environment for driver distraction: a thematic framework of sociotechnical factors. Appl. Ergon. 68, 213–228. Ponte, G., Baldock, M., Thompson, J., 2016. Examination of the Effectiveness and Acceptability of Mobile Phone Blocking Technology Among Drivers of Corporate Fleet Vehicles. Rahman, M.M., Lesch, M.F., Horrey, W.J., Strawderman, L., 2017. Assessing the utility of TAM, TPB, and UTAUT for advanced driver assistance systems. Accid. Anal. Prev. 108, 361–373. Reagan, I., Cicchino, R., 2018. Do Not Disturb While Driving: Use of Cellphone Blockers Among Adult Drivers. Insurance Institute for Highway Safety. Retrieved from https:// www.iihs.org/api/datastoredocument/bibliography/2179. Steinmetz, H., Knappstein, M., Ajzen, I., Schmidt, P., Kabst, R., 2016. How Effective are Behavior Change Interventions Based on the Theory of Planned Behavior? Z. Psychol. 224 (3), 216–233. Van Der Laan, J.D., Heino, A., De Waard, D., 1997. A simple procedure for the assessment of acceptance of advanced transport telematics. Transp. Res. Part C Emerg. Technol. 5 (1), 1–10. Venkatesh, V., Davis, F.D., 2000. A theoretical extension of the technology acceptance model: four longitudinal field studies. Manage. Sci. 46 (2), 186–204. Venkatesh, V., Morris, M.G., Davis, G.B., Davis, F.D., 2003. User acceptance of information technology: toward a unified view. Mis Q. 425–478. Visser, T.A., Bischof, W.F., Di Lollo, V., 2004. Rapid serial visual distraction: task-irrelevant items can produce an attentional blink. Percept. Psychophys. 66 (8), 1418–1432. Walsh, S.P., White, K.M., Hyde, M.K., Watson, B., 2008. Dialling and driving: factors influencing intentions to use a mobile phone while driving. Accid. Anal. Prev. 40 (6), 1893–1900. Yao, X., Wu, Y., Liu, H., Zhao, X., Bian, Y., Qu, W., 2019. Analysis of psychological influences on navigation use while driving based on extended theory of planned behavior. Transp. Res. Rec 0361198119845666. Young, K.L., Osborne, R., Koppel, S., Charlton, J.L., Grzebieta, R., Williamson, A., et al., 2019. What contextual and demographic factors predict drivers’ decision to engage in secondary tasks? IET Intell. Transp. Syst. 13 (8). https://doi.org/10.1049/iet-its. 2018.5546.

Acknowledgement Dr Oviedo-Trespalacios is the recipient of an Australian Research Council Discovery Early Career Researcher Award [DE200101079] funded by the Australian Government. References Adell, E., 2010. Acceptance of driver support systems. Paper Presented at the Proceedings of the European Conference on Human Centred Design for Intelligent Transport Systems. Ajzen, I., 1991. The theory of planned behavior. Organ. Behav. Hum. Decis. Process. 50 (2), 179–211. Albert, G., Lotan, T., 2019. Exploring the impact of“ soft blocking” on smartphone usage of young drivers. Accid. Anal. Prev. 125, 56–62. Buckley, L., Kaye, S.-A., Pradhan, A.K., 2018. Psychosocial factors associated with intended use of automated vehicles: a simulated driving study. Accid. Anal. Prev. 115, 202–208. Davis, F.D., 1985. A Technology Acceptance Model for Empirically Testing New End-user Information Systems: Theory and Results. Massachusetts Institute of Technology. Davis, F.D., 1989. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 319–340. De Angelis, M., Puchades, V.M., Fraboni, F., Pietrantoni, L., Prati, G., 2017. Negative attitudes towards cyclists influence acceptance of an in-vehicle cyclist detection system. Transp. Res. Part F Traffic Psychol. Behav. 49, 244–256. https://doi.org/10. 1016/j.trf.2017.06.021. Delgado, M.K., McDonald, C.C., Winston, F.K., Halpern, S.D., Buttenheim, A.M., Setubal, C., et al., 2018. Attitudes on technological, social, and behavioral economic strategies to reduce cellphone use among teens while driving. Traffic Inj. Prev. 1–8. Di Lollo, V., Kawahara, J.I., Ghorashi, S.S., Enns, J.T., 2005. The attentional blink: Resource depletion or temporary loss of control? Psychol. Res. 69 (3), 191–200. Dingus, T.A., Guo, F., Lee, S., Antin, J.F., Perez, M., Buchanan-King, M., Hankey, J., 2016. Driver crash risk factors and prevalence evaluation using naturalistic driving data. Proc. Natl. Acad. Sci. U. S. A. 113 (10), 2636–2641. Gauld, C.S., Lewis, I., White, K.M., Fleiter, J.J., Watson, B., 2017. Smartphone use while driving: what factors predict young drivers’ intentions to initiate, read, and respond to social interactive technology? Comput. Human Behav. 76, 174–183. https://doi. org/10.1016/j.chb.2017.07.023. Huemer, A.K., Schumacher, M., Mennecke, M., Vollrath, M., 2018. Systematic review of observational studies on secondary task engagement while driving. Accid. Anal. Prev. 119, 225–236. https://doi.org/10.1016/j.aap.2018.07.017. Larue, G.S., Rakotonirainy, A., Haworth, N.L., Darvell, M., 2015. Assessing driver acceptance of Intelligent Transport Systems in the context of railway level crossings. Transp. Res. Part F Traffic Psychol. Behav. 30, 1–13. Madigan, R., Louw, T., Wilbrink, M., Schieben, A., Merat, N., 2017. What influences the decision to use automated public transport? Using UTAUT to understand public acceptance of automated road transport systems. Transp. Res. Part F Traffic Psychol. Behav. 50, 55–64. McGinn, M.C., 2014. Predicting Factors for Use of Texting and Driving Applications and the Effect on Changing Behaviors. Southern Illinois University at Edwardsville. Monsell, S., 2003. Task switching. Trends Cogn. Sci. 7 (3), 134–140. National Highway Traffic Safety Administration, 2016. Visual-Manual NHTSA Driver Distraction Guidelines for Portable and Aftermarket Devices. National Highway Traffic Safety Administration (NHTSA), Department of Transportation (DOT), Washington, DC. Oviedo-Trespalacios, O., 2018. Getting away with texting: behavioural adaptation of drivers engaging in visual-manual tasks while driving. Transp. Res. Part A Policy Pract. 116, 112–121. https://doi.org/10.1016/j.tra.2018.05.006. Oviedo-Trespalacios, O., Haque, M.M., King, M., Washington, S., 2016. Understanding the impacts of mobile phone distraction on driving performance: a systematic review. Transp. Res. Part C Emerg. Technol. 72, 360–380. Oviedo-Trespalacios, O., King, M., Haque, M.M., Washington, S., 2017. Risk factors of

9